From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Brian Herlihy <btherl(at)yahoo(dot)com(dot)au> |
Cc: | pgsql-novice(at)postgresql(dot)org |
Subject: | Re: Efficiency of CREATE TABLE AS |
Date: | 2007-05-09 05:57:31 |
Message-ID: | 21701.1178690251@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-novice |
Brian Herlihy <btherl(at)yahoo(dot)com(dot)au> writes:
> If I do the following
> CREATE TABLE foo AS SELECT * FROM bar
> and a sequential scan is chosen as the plan (in particular, no
> ordering required), will the rows be written directly into the new
> table?
Yes ... actually, the plan has nothing to do with it.
> I am concerned about the case where there may be many
> gigabytes of data being copied between these two tables. And my
> precise concern is that the data may be stored temporarily before
> being placed in the new table, slowing things down.
No. The computed tuples go to the new table, noplace else. See
intorel_receive() and related code.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Loredana Curugiu | 2007-05-09 12:07:37 | Count rows group by time intervals |
Previous Message | Brian Herlihy | 2007-05-09 05:34:10 | Efficiency of CREATE TABLE AS |